The associations between multimorbidity, healthcare resource utilisation and costs in individuals with Alzheimer's disease dementia

Background

Multimorbidity (having two or more comorbidities), is common among people with dementia and associated with lower survival and increased healthcare use. We aimed to identify comorbidity clusters in people living with Alzheimer’s disease (AD) dementia in the UK, describe comorbidity-cluster prevalence and patterns, and estimate associations of clusters with healthcare resource utilisation (HCRU) and costs.

Methods

This was a cohort study using Discover dataset (electronic health records from approximately 2.8 million North-West London residents). We identified individuals with AD dementia using diagnostic codes, and estimated HCRU and total healthcare costs (including primary and hospital-based care). Individuals were grouped based on comorbidity profile using k-medoids clustering. Multivariable modelling was used to estimate associations between comorbidity clusters and healthcare costs.

Results

Among 18,116 individuals with AD dementia, eight comorbidity clusters were identified. The three highest-cost clusters incurred mean costs per patient year(ppy) of £6355, £5560, and £5284 respectively. The highest-cost cluster had a high burden of comorbidities with the most prevalent: frailty (83.2 %), hypertension (81.8 %), type 2 diabetes mellitus (81.4 %), and chronic kidney disease (69.7 %). Costs in the three highest-cost clusters were 1.7 to 2.0 times higher than the lowest-cost cluster (mean cost £3160ppy), which featured the lowest overall comorbidity burden. The lowest median survival times were in the two highest-cost clusters. Cluster membership was strongly associated with costs even after adjusting for a wide range of demographic and clinical factors. Compared with the lowest-cost cluster the average marginal increases in costs for the clusters ranged between £1072 (95 %CI:£478-£1666) and £3531ppy (95 %CI: £2850-£4212).

Conclusions

HCRU and costs in individuals with AD dementia show notable differences, with a large proportion of costs attributable to a minority of individuals with multiple comorbidities (particularly cardiometabolic diseases and frailty). Health systems should ensure initiatives to improve timely diagnosis and treatment of people with AD taking account of the high multimorbidity prevalence in this population.